Introduction

Top Abstract Introduction Experimental Procedures Results Discussion Appendix References

Estrone sulfate (E₁S¹) is one of the compounds used in hormone replacement therapy. The activity of E₁S results from the release of active estrone (E₁) through desulfation in liver, and E₁ can be sulfated back to the inactive E₁S. Typically, the rate of sulfation of E₁ is measured, but the observation may be erroneously based on the net rate of formation of E₁S. Since sulfoconjugation plays an important role in the deactivation and elimination of E₁, it is critical to consider the roles of sulfation and desulfation on the duration of activity of estrogens. When sulfation or desulfation is altered in disease states, the apparent formation of E₁S will be further modulated.

One of the key enzymes responsible for the sulfation of E₁ is estrogen sulfotransferase, a cytosolic enzyme that also sulfates estradiol (E₂) in the presence of the obligate cosubstrate, 3'-phosphoadenosine 5'-phosphosulfate (PAPS). Both human hydroxysteroid and phenol sulfotransferases are known to sulfate estrogens (Falany et al., 1994), but these possess lower affinities towards E₁ (Falany et al., 1994, 1995). Thus, under physiological conditions, E₁ is more likely to be sulfated predominantly by estrogen sulfotransferase than by other isoforms of sulfotransferases. Previous studies had revealed that estrogen sulfotransferase was localized more abundantly in the PV than in the PP region of the rat liver. On the other hand, hydroxysteroid sulfotransferase was predominantly present in the PP hepatocytes, and the phenol sulfotransferase was evenly distributed in the liver acinus (Tosh et al., 1996).

Estrone sulfatase, a membrane-bound enzyme with a pH optimum of 7.4 is mainly responsible for the desulfation of E₁S and exhibits its highest activity in the liver (Milewich et al., 1984). Estrone sulfatase is also known as arylsulfatase C, a microsomal enzyme that copurifies with steroid sulfatase and cleaves the sulfate moiety of several 3-hydroxysteroid sulfates (Dolly et al., 1972). In female rats, an abundance of estrone sulfatase activity was found in both the endoplasmic reticulum and nucleus (Zhu et al., 1998). Previous studies suggest that desulfation of 4-methylumbelliferyl sulfate by arylsulfatase C was homogeneous across the rat liver (Anundi et al., 1986) and human liver tissues (El Mouelhi and Kauffman, 1986).

The phenomenon of futile cycling between E₁ and E₁S has yet to be viewed in conjunction with hepatic transport of E₁ and E₁S and other metabolic pathways of E₁. Transport of E₁S in rat liver is found to be mediated by the organic anion-transporting polypeptides, Oatp1 (Jacquemin et al., 1994), Oatp2 (Noé et al., 1997), and Oatp4 (Cattori et al., 2000); the multispecific organic anion transporter 3, OAT3 (Kusuhara et al., 1999); and the sodium-dependent taurocholate-cotransporting polypeptide, Ntcp (Hagenbuch et al., 1991). The K_m values for the uptake of E₁S mediated by Oatp1 (Bossuyt et al., 1996), Oatp2 (Noé et al., 1997), and Ntcp (Schroéder et al., 1998) expressed in the Xenopus laevis oocytes were around 4.5 to 27 µM, showing that transport is of high affinity. Transport of E₁S was defined by these sinusoidal transporters and passive diffusion and was found to be similar among perivenous (PV) and periportal (PP) rat hepatocytes (Tan et al., 1999).

The objective of this study was to highlight the importance of transport, metabolism, and zonal aspects to understand their interplay on the futile cycling of estrogens in intact hepatocytes and, ultimately, the whole organ. In this article, metabolic activities in subcellular fractions of enriched PP and PV rat hepatocytes were assessed and in turn related to cellular expressions of rat liver estrone sulfatase and estrone sulfotransferase. These in vitro metabolic parameters and previously obtained transport parameters on hepatocyte uptake (Tan et al., 1999) were then used to describe the futile cycling between E₁S and E₁ in intact cells when E₁S (1, 5, 25, and 125 µM) was incubated with PP and PV hepatocytes.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion Appendix References

Materials. [6,7-³H]E₁S (ammonium salt; specific activity, 53 Ci/mmol), [6,7-³H]E₁ (specific activity, 40.6 Ci/mmol), and [4-¹⁴C]E₁ (specific activity, 56.6 Ci/mol) were purchased from NEN Life Science Products (Boston, MA). The radiochemical purities, as found by high-performance liquid chromatography (HPLC) or thin-layer chromatography, were greater than 95%. Unlabeled E₁S, E₁, PAPS, and bovine serum albumin (fraction V) were purchased from Sigma Chemical Co. (St. Louis, MO). Collagenase was obtained from Boehringer Mannheim (Oakville, ON, Canada). Digitonin was acquired from Fluka Chemie (Buchs, Switzerland). Antisera raised against rat liver phenol sulfotransferase (rSULT1A1), hydroxysteroid sulfotransferase (rSULT2A1), and estrogen sulfotransferase (rSULT1E1) were kindly provided by Dr. Charles N. Falany (University of Alabama, Birmingham, AL). The immunoblot Assay Kit (mini-Protean II systems) was obtained from Bio-Rad Laboratories (Mississauga, ON, Canada). Goat anti-rabbit horseradish peroxidase conjugate was obtained from Amersham Pharmacia Biotech (Oakville, ON, Canada). All other reagents were of the highest available grade.

Isolation of Zonal Rat Hepatocytes, Lysates, and Subcellular Fractions of Zonal Hepatocytes. Male Sprague-Dawley rats (275-325 g, Charles River Canada, St. Constant, QC, Canada) were used for the preparation of isolated hepatocytes. Rats were housed in accordance with Animal Protocols at the University of Toronto under a 12:12-h light:dark cycle and were given food and water ad libitum.

Estrone Sulfatase Activity. Estrone sulfatase activity in the S9 of zonal hepatocytes was determined by formation of [³H]E₁ from [³H]E₁S. After preincubation of zonal S9 and E₁S separately at 37°C for 5 min, the solutions were combined to result in a mixture of S9 protein (1.4 mg) and E₁S (1-200 µM with 1-177 × 10⁵ dpm/ml of [³H]E₁S) in a final volume of 0.4 ml of Tris-HCl buffer (25 mM) at pH 7.4. Samples (0.1 ml) were then removed at 2 min, a predetermined time in which product formation was linear with time, into 0.5 ml of acetonitrile containing 4 µM danazol, the internal standard for HPLC analyses.

E₁ Sulfotransferase Activity. Estrone sulfotransferase activities in the zonal lysates and cytosolic fractions were estimated from that rates of formation of [³H]E₁S from [³H]E₁. Lysates and cytosolic fractions of zonal hepatocytes, which were preincubated at 37°C for 5 min, were added to mixtures of PAPS, E₁, and [³H]E₁ (2.0 ± 0.1 × 10⁵ dpm/ml) to result in 1 µM E₁ and 700 µM PAPS in 1 ml of Tris-HCl buffer (25 mM) at pH 7.4. To ensure sufficient cosubstrate for the reaction, the PAPS concentration chosen was higher than that reported for the rat liver (70 nmol/g of liver or 117 µM PAPS in cell water) (Brzeznicka et al., 1987). Samples (0.2 ml) were removed at 6 min, a predetermined time in which E₁ sulfation was linear with time. The samples were added to 0.8 ml of acetonitrile containing 4 µM danazol.

Metabolism of E₁S in Intact Zonal Hepatocytes. Zonal hepatocyte suspensions (2 × 10⁶ cells/ml), preincubated for 10 min at 37°C in the incubation buffer, were added to equivolumes of unlabeled E₁S and [³H]E₁S (5.1 ± 0.3 × 10⁶ dpm/ml) prepared in incubation buffer to result in 1, 5, 25, and 125 µM E₁S in 10⁶ cells/ml. Two samples were retrieved at various times (1-30 min) from the incubation mixture. The first sample (100 µl) was deproteinized immediately with 0.4 ml of acetonitrile containing 4 µM danazol, and the second sample (150 µl) was placed immediately into a polyethylene microfuge tube (300 µl) containing 100 µl of 1-bromododecane. Upon centrifugation at 9000g (2 s), hepatocytes were removed into the layer of 1-bromododecane, and 100 µl of the resultant extracellular medium remaining on top was removed into a 1.5-ml microcentrifuge tube containing 4 µM danazol in 0.4 ml of acetonitrile. Subsequently, the contents of E₁ and E₁S in the incubation mixture and extracellular medium were assayed by HPLC. The amount of E₁S or E₁ in the cellular space was estimated by the difference of the quantities in the incubation mixture and extracellular medium. The difference in mass between the administered amount and the sum of E₁S and E₁ provided an estimate of the formation of other E₁ metabolites (M; estrone glucuronide or E₁G, estradiol or E₂, and its conjugates) at various times.

Protein Binding/Metabolism in Extracellular Medium. During the preparation of isolated hepatocytes, protein debris from suspending dead cells (0.16 ± 0.06 mg/10⁶ cells) was found to persist routinely in the incubation mixture despite the washings; centrifugation with percoll failed to remove the presence of the protein debris fragments. In view of the tight binding of E₁S and E₁ to albumin (Rosenthal et al., 1972; Rao, 1998), binding of E₁S and E₁ to the protein debris in the extracellular medium was estimated by ultrafiltration (Centricon 3, Amicon Inc., MA). Solutions containing [³H]E₁ (8.5 × 10⁵ dpm/ml) or [³H]E₁S (4.9 ± 2.4 × 10⁵ dpm/ml) plus E₁S (0.8-250 µM) were added to blank extracellular medium of the incubation mixture, which was prepared in the absence of drug. After incubation of the mixture for 10 min at 37°C, an aliquot (2 ml) was removed into a Centricon tube and centrifuged at 2500g for 20 min at 37°C. Liquid scintillation fluor (5 ml, Ready Safe, Beckman Coulter, Mississauga, ON, Canada) was added to the original mixture (0.2 ml) and the resulting filtrate (0.2 ml) in different minicounting vials, then quantified by liquid scintillation spectrometry (model LS6800, Beckman Coulter). Leakage of protein through the Centricon filter was less than 0.5% of the mixture solution. Desulfation of E₁S and metabolism of E₁ within the extracellular medium were less than 1% over the experimental time and were deemed insignificant.

Immunoblot Analysis. Lysates, centrifuged at 100,000g for 60 min at 4°C, and the cytosolic fraction of zonal hepatocytes, prepared as described by Tirona et al. (1999), was used for immunoblot analysis. Aliquots containing 10 µg of protein were resolved by SDS-PAGE in a 12% polyacrylamide gel and electrophoretically transferred to nitrocellulose membrane. Primary rabbit antibodies (anti-rSULT1A1 at 1:20,000, anti-rSULT2A1 at 1:10,000, and anti-rSULT1E1 at 1:20,000 dilution) were then incubated with the blots for 1 h at room temperature. Finally, goat anti-rabbit IgG horseradish peroxidase conjugate (1:40,000) was used as the secondary antibody, and the immunoconjugates were detected by chemiluminescence (Amersham). The intensity of the protein band was integrated using the NIH Image software (http://rsb.info.nih.gov/).

Protein Assay. In all preparations, protein was determined by the method of Lowry et al. (1951), with bovine serum albumin as the standard.

HPLC Analysis. Liquid chromatography was performed on a Shimadzu HPLC system (Shimadzu Corporation, Kyoto, Japan) equipped with a SCL-10A system controller, a SPD-10A UV-visible detector, a LC-10AT solvent delivery system, a SIL-10A-XL automatic injector, and a CR-4A chromatopac integrator. Separation was carried out by a 10-µm µBondapak C₁₈ reversed-phase column (39 cm × 300-mm i.d.; Waters, Milford, MA). A binary gradient consisting of mobile phase A (10 mM ammonium acetate, pH 7.5) and mobile phase B (acetonitrile) was used at a constant flow rate of 1 ml/min for the separation of E₁G, E₁S, E₁, E₂, and danazol. Initially, the wavelength of detection was set at 270 nm for the detection of E₁G and E₁S, and the mobile phase was linearly increased from 10% B to 50% B in 15 min. Then the mobile phase was maintained at 50% B for 10 min, and the wavelength was altered to 285 nm at 20 min for the detection of E₁ and E₂. At 25 min, the mobile phase was linearly increased to 75% B in 2 min and was maintained for another 10 min before being brought back down to 10% B over the next 5 min. The mobile phase was maintained at 10% B for an additional 5 min to re-equilibrate the column. The retention times were as follows: E₁G, 14 min; E₁S, 16 min; E₁, 24 min; E₂, 26 min; and danazol, 34 min.

Kinetic Analysis for Extracellular Binding of E₁S and E₁. The binding of E₁S and E₁ to proteins that were present in the extracellular medium was described by the Langmuir binding isotherm (eq. 1), where the bound concentration of extracellular E₁S ([E₁S_bound]_ec) was related to the unbound concentration ([E₁S_unbound]_ec) by the binding dissociation constant, K, and the product of number of binding site (n^E₁S) and the total protein concentration in the extracellular medium ([P_total]_ec). The kinetic constants were obtained by regression of [E₁S_bound]_ec versus [E₁S_unbound]_ec according to eq. 1 with use of the fitting software package SCIENTIST (version 2; MicroMath Scientific Software, Salt Lake City, UT) and an optimized weighting scheme (1/observation²):

(1)

After obtaining K and n^E₁S[P_total]_ec from eq. 1, eqs. 2 and 3 were used to estimate the tissue binding of E₁S in the hepatocyte from known total protein concentration in cell-homogenate [P_total]_c.

(2)

(3)

(4)

Kinetic Modeling of E₁S and E₁ Disposition in Intact Zonal Hepatocytes. A cellular, kinetic model that considered protein binding in both extracellular and cellular spaces, transport, metabolism, interconversion, and an intracellular vesicular compartment (Fig. 1) best described the concentration- and time-dependent data of E₁S and E₁ in enriched PP and PV hepatocytes. Inclusion of binding in both cellular and extracellular medium is justified in view of the demonstrable binding but lack of metabolism in extracellular medium, which contained protein debris from cells. E₁S binding and debinding to protein are denoted by on- and off-rate constants, k and k, respectively; these constants and the ratio K (k/k) are expected to be identical for both extracellular and cell media. However, the protein concentration or [P_total]_ec in the extracellular medium is only a fraction of that in cells since the [P_total]_c is much higher. The effective binding concentration in cell was determined by multiplication of the ratio of protein concentrations in cell/debris ([P_total]_c/[P_total]_ec) with n^E₁S[P_total]_ec according to eq. 2. For the sake of simplicity, the unbound concentration of E₁ is represented by the product of the unbound fractions of E₁ (f and f represent the unbound fraction of E₁ in the cell and the extracellular medium, respectively) and the total concentration of E₁. Since the canalicular membrane on the cell surface that mediates excretion of anions is known to internalize to form vesicles shortly after hepatocyte isolation (Oude Elferink et al., 1993), a similar phenomenon was postulated to exist for E₁S accumulation into vesicles. A vesicular compartment is included for E₁S since E₁S undergoes demonstrable biliary excretion in the perfused rat liver (Tan et al., 2001).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   A cellular kinetic model for the futile cycling of E1 and E1S in intact isolated hepatocytes. Cellular uptake of E1S consisted of a saturable and a bidirectional transport system, whereas E1 transport was described by a bidirectional transmembrane clearance. From the preparative procedure, membranous debris capable of binding to E1 and E1S remained in the extracellular medium. Binding on and off rate constants were used to describe the nonlinear binding of E1S, whereas a constant, unbound fraction was assumed for E1 binding. For the range of E1S concentrations (1, 5, 25, and 125 µM), desulfation, sulfation, and formation of other estrone metabolites (M) are pathways best described by saturable systems. Internalization of canalicular membrane from the hepatocyte surface to vesicle was assumed to occur shortly after isolation of the hepatocytes into a vesicular compartment for E1S. See text for the detailed description of the terms.

Fitting. Mass balanced rate equations (see Appendix) were written to describe events of the cellular kinetic model (Fig. 1). Fitting was performed by the software package SCIENTIST based on experimentally obtained binding, metabolic, and transport parameters (Table 1). The parameterstransport clearance of E₁ (P), the Michaelis-Menten constant for uptake of E₁S into hepatocytes (K), the vesicular intrinsic clearance for accumulation of E₁S (CL), the pooled kinetic constants for formation of the composite of E₁ metabolites M (V and K), the kinetic constants for E₁ sulfation (V and K, the latter constrained as V/CL), and the tissue binding of E₁were optimized by least-square fitting with appropriate weighting schemes of 1/observation (for data increasing in value) and 1/observation² (for data decreasing in value). The goodness of fit was viewed with respect to the coefficient of variation (standard deviation of parameter estimate/parameter value), the residual plot, and the Model Selection Criterion (MSC).

Statistics. All data were presented as the mean ± S.D., and the means were compared by use of ANOVA, with the P value of 0.05 as the level of significance. A paired t test was used to compare the means for the data on zonal lysates since the same liver was used for the preparation of both PP and PV lysates; the level of significance was set at 0.05.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion Appendix References

Biochemical Characterization of Zonal Hepatocytes and Lysates. The marker enzymes, alanine aminotransferase and glutamine synthetase, verified that enriched PP and PV hepatocytes and lysates were isolated from different acinar regions of the rat liver. The PP/PV ratios of the marker enzyme alanine aminotransferase for zonal hepatocytes and zonal lysates were 1.9 ± 0.9 and 7.4 ± 4.2, respectively. The acinar gradient of alanine aminotransferase content in zonal lysates was steeper because the preparations were obtained from the most distal and proximal acinar regions of the rat liver. However, the PP/PV ratios of the marker enzyme glutamine synthetase for the zonal hepatocytes and lysates were 0.027 ± 0.023 and 0.029 ± 0.013, respectively, and were similar. These PP/PV ratios were in agreement with those reported by others (Lindros and Penttïla, 1985; Quistorff and Grunnet, 1987; Tosh et al., 1996; Tirona et al., 1999).

Desulfation of E₁S in the Zonal S9 Preparations. Preliminary study failed to show E₁ sulfation in absence of PAPS within the zonal S9 preparations. Therefore, the rate of E₁ formation from E₁S in S9 represented the true desulfation rate. The results were best fit to the Michaelis-Menten equation (Table 1 and Fig. 2), yielding similar K (30 to 34 µM) and V (2.4-2.9 nmol/min/S9 protein) values for estrone desulfation in the PP and the PV preparations (ANOVA, P > 0.05). The K was much lower in relation to that reported for another well studied sulfate conjugate, 4-methylumbelliferyl sulfate (Chiba and Pang, 1993).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Rates of desulfation (mean ± S.D.) for E1S in the S9 fractions of the PP and PV hepatocytes at 37°C (n = 4). Lines, fitted lines based on the Michaelis-Menten equation with an optimal weighting scheme of unity.

E₁ Sulfation in Zonal Lysates and the Cytosol of Zonal Hepatocytes. Estrone sulfotransferase activity was detected in zonal lysates and cytosolic fractions of PP and PV hepatocytes in the presence of PAPS. Desulfation of E₁S in zonal lysate and zonal cytosol was less than 1% over the experimental time and was negligible. Hence in both lysate and cytosol, the observations reflected the true and not the net sulfation activity of E₁ sulfotransferases. The activity was significantly higher in the PV region than the PP region (Table 2). The PV/PP ratios of estrone sulfation activities in the cytosolic fractions and lysates were 4.3 ± 2.2 and 3.1 ± 1.8, respectively.

Immunoblot of rSULT1A1, rSULT2A1, and rSULT1E1 in Zonal Hepatocytes and Zonal Lysates. The rSULT1E1 protein in the cytosolic fraction of PV hepatocytes was 3.4 ± 1.1 times that of PP and paralleled the results obtained previously for CYP1A2, a PV marker within the same cell preparations (PV/PP ratio of 4.1 ± 3.3; Tirona et al., 1999). On the other hand, the rSULT2A1 protein in PP hepatocytes was 3.5 ± 2.7 times that of the PV hepatocytes, and the rSULT1A1 protein was not significantly different (ANOVA, P > 0.05; PV/PP ratio of 1.13 ± 0.23) among zonal hepatocytes (Fig. 3). Trends for the zonal lysates remained similar to those of the zonal hepatocytes: the rSULT1E1 protein in the PV lysates was 4.0 ± 3.0 times that of the PP lysates. The rSULT2A1 protein was exclusively found in the PP lysates, and the rSULT1A1 protein was again present evenly among zonal lysates (Fig. 4).

View larger version (48K): [in this window] [in a new window]   Fig. 3.   Immunoblot analysis of rSULT1A1, rSULT2A1, and rSULT1E1. Cytosol proteins (9 µg), obtained from PV and PP isolated hepatocytes, were separated by SDS-PAGE in 12% polyacrylamide gels and probed with rabbit anti-rSULT1A1, anti-rSULT2A1, and anti-rSULT1E1 as described under Experimental Procedures. CYP1A2, a PV marker enzyme in the same preparations, verified the enrichment of PP and PV hepatocytes. The zonal crude membrane proteins (100,000g pellets, 10 µg) were resolved by SDS-PAGE in 9% polyacrylamide gels and probed with anti-rat CYP1A2 as described by Tirona et al. (1999). * The quantity of immunoreactive protein present in PV hepatocytes was significantly different from that of the PP hepatocytes (ANOVA, P < 0.05).

View larger version (40K): [in this window] [in a new window]   Fig. 4.   Immunoblot analysis of zonal cell lysates with anti-rSULT1A1, anti-rSULT2A1, and anti-rSULT1E1. Proteins (6.6 µg) from zonal lysates were resolved by SDS-PAGE in 12% polyacrylamide gels as described under Experimental Procedures. Lanes 1, 2, 3, and 4 represent paired periportal lysate (LPP) and perivenous lysate (LPV) derived from four rat livers. ** The quantity of immunoreactive protein in perivenous lysate was significantly different from that of periportal lysate (paired t test, P < 0.05).

Protein Binding of E₁S and E₁. The binding data of E₁S (0.8-250 µM) in extracellular medium were best described by the Langmuir binding isotherm (eq. 1). Assuming that the molecular mass of binding proteins were around 82 kDa (the molecular mass of estrogen binding protein; Rao, 1998), the n^E₁S[P_total]_ec and the K for E₁S were estimated to be 23.5 ± 3.0 and 23.4 ± 4.5 µM, respectively, and the number of binding sites (n^E₁S) was 12 ± 1.5, based on the measured extracellular protein concentration of 0.16 ± 0.03 mg/ml (Fig. 5A). The unbound fractions of E₁S at 1 and 125 µM in the extracellular medium were around 0.5 and 0.8, respectively. In tissue, where the cellular protein concentration (1.6 ± 0.3 mg/ml) was about 10 times the extracellular concentration, the unbound fractions of E₁S in tissues were calculated according to eq. 3, based on the K and n^E₁S. The unbound fractions of E₁S in tissue were predicted to vary from 0.1 to 0.7, as shown in Fig. 5B.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Protein binding of E1S. A, protein binding profile of E1S (0.8-250 µM; mean ± S.D.) in extracellular medium. The line represents the predictions based on the fitted parameters with eq. 1 from an optimal weighting scheme of 1/observation2. B, the unbound fraction of E1S in extracellular medium () and the predicted unbound fraction of E1S in tissue (-) versus the total concentrations of E1S.

Metabolism of E₁S in Intact Zonal Hepatocytes. Concentration-dependent metabolism of E₁S was observed in intact PP and PV hepatocytes (Fig. 6A). For both PP and PV hepatocyte preparations, biphasic elimination patterns were observed for the lower concentrations of E₁S (<25 µM), whereas monoexponential decay profiles were observed at the highest concentration of E₁S used (125 µM). The patterns of E₁S in extracellular medium paralleled those in the incubation mixture (Fig. 6B) and were similar for both PP and PV cells. However, the pattern differed dramatically in the cell wherein cellular concentrations of E₁S were much higher than those extracellularly (Fig. 6C), yielding apparent tissue to medium partitioning ratios of greater than unity in both PP and PV cells (Fig. 7A). The apparent partition coefficients of E₁S at equilibrium decreased with increasing E₁S concentration and were similar for both PP and PV cells (Fig. 7B). As shown in Table 3, nonlinear kinetics were shown to exist with increasing E₁S doses. Values of the AUC of E₁S were higher in PP than in PV hepatocytes, although statistical significance was not found (Table 3). The apparent clearance of E₁S [dose/AUC(0  )] decreased with increasing dose and was lower in PP hepatocytes than in PV hepatocytes. Again, statistical difference was not found due to the large interanimal variability.

View larger version (36K): [in this window] [in a new window]   Fig. 6.   Time- and concentration-dependent profiles for metabolism of E1S and E1 in the isolated PP and PV hepatocytes at 37°C (n = 5). A, total concentrations of E1S in incubation mixture; B, extracellular concentrations of E1S; C, cellular concentrations of E1S; D, total concentrations of E1 in incubation mixture; E, cellular concentrations of E1; and F, extracellular concentrations of E1. The concentrations of E1S or E1 associated with the four numbered initial concentrations of E1S (1, 5, 25, and 125 µM) are shown. Observations for 1 (), 5 (), 25 (), and 125 µM () E1S with PP hepatocytes were compared with corresponding observations with PV hepatocytes (open symbols). () and (), fitted lines for the PP and PV preparations, respectively, based on eqs. A1 to A11 with appropriate weighting schemes of 1/observation and 1/observation2. There was a general failure to detect data points difference between PP and PV due to the large variability existing among the animal preparations. * Data for PP hepatocytes that were significantly different from the corresponding timed data of PV hepatocytes (ANOVA, P < 0.05).

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Partitioning of E1S between cell and medium in PP and PV hepatocyte systems. A, time-dependent profiles for the partition coefficients of E1S, cellular concentration/extracellular concentration, in relation to the different initial concentrations of E1S [1 (), 5 (), 25 (), and 125 µM (); mean ± S.D.], and the corresponding open symbols represent the PV data. B, a decreasing pattern of the partition coefficients of E1S existed at equilibrium for the various initial concentrations of E1S (1-125 µM).

Fitted Results for E₁S and E₁ in Intact Zonal Hepatocytes with the Kinetic Model. When simultaneous fitting was performed on the total, extracellular, and cellular E₁S and E₁ data for each set of experiments consisting of four E₁S initial concentrations and the same pool of hepatocytes, good fits were obtained, although high coefficients of variation were found associated with the fitted parameters. The optimized fit that considered tissue binding and vesicular storage of E₁S is presented in Fig. 6, and the mean of the optimized parameters of five experiments and the assigned parameters are summarized in Table 4.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion Appendix References

Estrone sulfate plays a vital role in the cycling of estrogens. Being hydrophilic, E₁S serves as a mobile estrogen and allows easy delivery to target tissues. E₁S, a common substrate of Oatp1, Oatp2, Oatp4, Ntcp, and OAT3, gains ready access into the liver tissue where it is desulfated to yield the active E₁. The cycling between E₁S and E₁ is influenced not only by metabolic activities on desulfation and sulfation, but also by the presence of other competitive pathways for E₁ metabolism, excretion of E₁S, the transmembrane characteristics, and tissue protein binding, recognizing that some of the proteins mediating the processes may be zonated in the acinus. Hence, we assessed the zonal, metabolic activities in tissue and integrated these with the transport and binding activities to examine the influence of metabolic heterogeneity on the futile cycling of estrogens in intact zonal hepatocytes. Notably, in contrast to parallel decay profiles in both extracellular medium for both parent and metabolite (Ebling and Jusko, 1986), we observed different decay half-lives for E₁S and E₁ in the hepatocyte system (Fig. 6).

From in vitro values of the kinetic constants for E₁S desulfation (K of 30 and 35 µM and V values of 2.2 and 1.9 nmol/min/10⁶ cells, respectively, for PP and PV hepatocytes), there was no difference in metabolic activity for both the proximal and distal regions of the rat liver. The observation was in good agreement with other findings on the homogeneous distribution of arylsulfatase C activity (Anundi et al., 1986). The K_m values were similar to that of a previous study (32 µM; Iwamori et al., 1976). By contrast, estrone sulfation was of higher affinity (K of 4.4-5.9 µM) in both PP and PV regions, but the V was much lower in value and differed between PP and PV hepatocytes (0.014-0.077 nmol/min/10⁶ cells, respectively; Table 4). The observation was consistent with the trends on sulfation of tracer estrone in both hepatocytes and lysates (Table 2) as well as with immunoblot analyses of rSULT1E1 (Figs. 3 and 4). The PP/PV ratio of rSULT1E1 protein was consistent with those reported by others (Tosh et al., 1996). But the low PV/PP ratio of rSULT2A1 was opposite to the observation on sulfation activities, suggesting that hydroxysteroid sulfotransferase contributes little to estrone sulfation. The even distribution of rSULT1A1 protein in the zonal cells also indicates that phenol sulfotransferase only plays a minor role in estrone sulfation. This evidence confirms that sulfation of estrone is predominantly catalyzed by estrogen sulfotransferase (rSULT1E1) in the presence of PAPS.

The CL was about 4 to 23 times higher than the CL (Table 5), and E₁ sulfation was the rate-limiting step in the futile cycling. In addition to E₁ sulfation, E₁ was metabolized to M with a much higher intrinsic clearance (CL; 328 and 495 µl/min/10⁶ cells, respectively, for PP and PV cells; Table 5) that showed a PV preponderance. The competitive metabolism of E₁ represents both glucuronidation by the UDP-glucuronosyltransferases that are localized pericentrally (Tosh and Burchill, 1996) and oxidation of E₁ by CYP1A2 and -3A that are concentrated in the PV region (Oinonen et al., 1996). This "pooled" CL was 38 to 106 times higher than the CL. Consequently, little E₁ is resulfated back to form E₁S. The higher activities for E₁ sulfation and formation of M in the PV region translates to the higher accumulation of E₁ in PP cells, as observed under low concentrations (cf. AUC values in Table 3).

Upon comparison of the metabolic intrinsic clearances of E₁ sulfation and E₁S desulfation to those for transport, the hepatic uptake clearances greatly exceed the metabolic intrinsic clearances (Table 5). The transport clearance of E₁S is rapid, but that for E₁ is even faster. The CL (Table 5) is substantial. Under physiological and first-order conditions where both E₁ and E₁S exist in low concentrations (nM), transport should remain very rapid and unsaturated. At high concentrations of E₁S, however, transport may become saturated at concentrations comparable with or exceeding K. The value of the fitted K is within the range of the K_m values (4.5-27 µM) reported for the various transporters and was similar to the value of K (24 µM) obtained in vitro (Tan et al., 1999). Adoption of the in vitro K value (24 µM), however, provided poorer fits. We found that the parameters for the transport systems of E₁S obtained from fitting were similar for both PP and PV hepatocytes, and the finding suggests the uniform distribution of transporters in rat liver. Uniform acinar distributions were found for Ntcp (Stieger et al., 1994), Oatp1 (Abu-Zahra et al., 2000), and Oatp2 (Tirona et al., 2000) in rat liver, and uptake of E₁S was similar in zonal hepatocytes (Tan et al., 1999). Saturation in uptake had occurred within the concentration range studied in the hepatocyte system, and this was shown by the decreasing partition coefficients of E₁S with increasing concentrations (Fig. 7B). Consistent with lack of zonation in uptake, values of the equilibrium partition coefficients of E₁S were similar for both PP and PV hepatocytes.

Although previous evidence has suggested that transport of E₁ across the membrane might involve carriers (Rao et al., 1977), our data were consistent with a linear, transmembrane flux for E₁ (P). The bidirectional uptake clearance for E₁ (1463-1484 µl/min/10⁶ cells) was even faster than that for E₁S, and no difference was found among PV and PP hepatocytes. The rapid transport clearance of E₁ was congruent with parallel trends of E₁ in cellular and extracellular medium that suggest rapid equilibration (Fig. 6, E and F). The high P is probably due to the high lipophilicity of E₁; indeed, the octanol/water log P value of E₁ is 3.1 (Howard and Meylan, 1997).

This study is the first account on binding of substrates to cell debris that resulted during hepatocyte preparation. The presence of extracellular binding of E₁S and E₁ has led to the conclusion that an even tighter tissue binding exists (Fig. 5B). Extracellular binding would decrease the uptake of E₁S and E₁, whereas cellular binding of E₁S and E₁ entraps the species within the cell and impedes cellular elimination. Tissue binding of E₁S and E₁ therefore exerts an important influence on the cellular kinetics of futile cycling of estrogens. Another issue that needs to be addressed with respect to tissue binding and metabolism is nonlinear tissue binding of E₁S and K < K (24 vs. 30-34 µM). The comparison of K_m values suggests that, with increasing cellular concentrations of E₁S, saturation of tissue binding precedes the saturation of the metabolic enzymes for desulfation. A similar scenariowith K_D for vascular binding of a flow-limited substrate < the K_mhad resulted in nonlinearity in drug clearance (Chiba and Pang, 1993; Xu et al., 1993). The same consequence will result here with nonlinearity in tissue binding.

To understand the interplay among the nonlinearity in transport, tissue binding, and the presence of vesicular accumulation of E₁S on the different t_1/2 values of E₁S and of E₁, simulations were further performed with the fitted parameters, with the substitution a single, nonsaturable uptake clearance of E₁S (CL of 246 µl/min/10⁶ cells), then a 10× higher dissociation binding constant (K was increased to 230 µM), and ultimately an absence of vesicular accumulation of E₁S. When only linear transport was introduced, linear decay of extracellular E₁S was observed. But the difference in t_1/2 values of E₁S and E₁ persisted (data not shown). The similarity in decay t_1/2 values of E₁S and of E₁ for any given dose could only be attained when CL was high and linear (246 µl/min/10⁶ cells), when K greatly exceeded K, and when there was a lack of vesicular accumulation of E₁S. The extracellular and cellular E₁S contents would now decay in unison with those of E₁, and similar half-lives were attained for both drug and metabolite species, as expected of the futile cycling phenomenon (Fig. 8). The pattern conforms to other reversible metabolic systems that describe the futile cycling between methylprednisolone and methylprednisone for which similar in vivo elimination half-lives were observed for both drug and metabolite (Ebling and Jusko, 1986). It may be thus concluded that the nonlinearity in uptake and tissue binding, and the presence of vesicular accumulation of E₁S, had resulted in different decay half-lives for E₁S and E₁ in the hepatocyte system.

View larger version (23K): [in this window] [in a new window]   Fig. 8.   Simulated profiles of E1S and E1 in the hepatocyte incubation system. Profiles were based on the PP parameters and substitution of a single, linear transport clearance of E1S (CL was set constant at 246 µl/min/106 cells), then a 10× higher binding dissociation constant (K was increased to 230 µM), and absence of vesicular accumulation of E1S. The concentrations of E1S or E1 associated with the four numbered initial concentrations of E1S (1, 5, 25, and 125 µM) are shown. A, total concentrations of E1S; B, extracellular concentrations of E1S; C, cellular concentrations of E1S; D, total concentrations of E1; E, cellular concentrations of E1; and F, extracellular concentrations of E1 were presented. The simulated lines for 1 (), 5 (), 25 (-), and 125 µM (······) of E1S were based on eqs. A1 to A11 (Appendix).

In conclusion, both E₁ and E₁S are rapidly taken up evenly into rat zonal hepatocytes. The sulfation of E₁ by estrogen sulfotransferase and the metabolism of estrone to other metabolites were more abundant in PV than in PP hepatocytes, although the desulfation of E₁S was evenly distributed. The rate-limiting factor for the futile cycling of E₁S and E₁ was sulfation, since transport was rapid and the intrinsic clearance of E₁S desulfation was higher than that of E₁ sulfation. The higher levels of E₁ and E₁S in PP hepatocytes were due to the higher PV metabolic activity towards E₁ sulfation and the formation of other metabolites. Different decay half-lives for E₁ and E₁S were observed, which were attributable to nonlinear uptake, tissue binding, and vesicular uptake of E₁S in the cell.